CIS 530: Artificial Intelligence

Please read the deep mind paper "Human-level control through deep reinforcement learning": http://www.nature.com/nature/journal/v518/n7540/full/nature14236.html

mnih2015.pdf

https://www.youtube.com/watch?v=He4t7Zekob0

https://youtu.be/V1eYniJ0Rnk

https://www.youtube.com/watch?v=Ih8EfvOzBOY

https://www.youtube.com/watch?v=IFmj5M5Q5jg

https://www.youtube.com/watch?v=M-QUkgk3HyE

This is a real useful book for building strong foundations in neural networks. We have covered chapter 2 and chapter 6 of the book.

\\172.30.10.2\FacultyShare\Fayyaz ul Amir Afsar Dr\CIS621 Machine Learning\Books\Laurene_Fausett_-_Fundamentals_of_neural_networks.PDF

-F

Dear Students,

All of you did really well in your projects. Here are a few minor things left in your projects. I would like you to cover them and upload your final project before 0900HRS on Wednesday, August 22, 2017 so that I can submit your final project grades. 

• Neural Style Transfer: Apply Neural Style Transfer for 10 paintings by Pakistani artists on 10 images of your choice.
• Visual Object Recognition: Try removing the delay in the system and make a short video recording of your system in action so we can show it off to others.
• OCR: Calculate the accuracy of the OCR system for printed text using a few documents for which the original text is also available.
• Sentiment Analysis: Try downloading a larger data set of tweets for analysis.
• Hemometer: Aqsa -- Normalization and showing the images to a doctor to identify whether the person has anemia or not. Amina -- Plot the leave one out cross-validation scores (so we can see why the AUC-PR is low).

Submission: All of you are to upload your class project data to piazza via private message to the instructor as follows:

Make a single zipped file for your project and it should contain three folders:

1. Data -- contains any and all data used in your project together with a readme.txt file explaining what the data files contain and how to use them. If your data set is large (videos!), you can give it to me via a flash disk.
2. Code -- contains any and all code used in your project together with a readme.txt file explaining how to get your code to run, any dependencies and installation requirements that it may have. The code should have 
3. Documentation -- contains any and all documents that you have developed for your project such as presentations containing proper references to all relevant documents. You can also upload papers relevant to your projects.

We shall have our presentations starting at 820am in B-215. Please be on time.

Each student shall have ~8 minute presentation, ~ 2 minutes Q/A session and ~5 minutes demo.

-F

The teaching staff has posted a new lecture notes resource.

Title: Reinforcement Learning
http://www.piazza.com/class_profile/get_resource/j4k7qqfhjfl5fd/j6ezfxwz1585bo

Lecture date: Aug 16, 2017

You can view it on the course page: https://piazza.com/pieas.edu.pk/summer2017/cis530/resources

We covered 3 basic rules of probability theory. I summarize them below:

• The probability is always [0,1], i.e., $0\le p(A)\le 1$
• If two events are independent then the probability of them happening together is the product of their individual probability values, i.e., $p(A,B)=p(A\cap B)=p(A)p(B)$
• If two events cannot happen together, then their probability that either of them happens is the sum of their individual probability values. $p(A\cup B)=p\left(A\right)+p\left(B\right)$. In case of conditional probability, i.e.,$p\left(A|B\right)=p(A,B)/p(B)$.